How to read SchoolData.ai
Methodology
SchoolData.ai is built on a simple premise: publish the facts, the whole facts, and nothing but the facts. We do not calculate proprietary school scores, assign letter grades, or claim that state assessments are perfectly comparable across state lines.
1. Verbatim Public Records
We directly ingest data from federal and state agencies. When a data point is suppressed for privacy reasons (e.g., small group sizes), we show it as suppressed. When data is missing from the public record, we show it as missing. We do not impute or guess values to create a "cleaner" visual interface.
2. State-Scoped Comparisons
Every state designs its own standardized tests and determines its own proficiency cut-scores. Therefore, a "proficient" student in one state may not meet the same bar in another state. SchoolData.ai strictly scopes assessment comparisons within state lines to prevent misleading national rankings.
3. Neighborhood Context is District Context
We map Census Bureau estimates (such as household income and broadband access) onto school district boundaries using the EDGE geocoding dataset. This allows users to understand the socioeconomic factors that surround a school, distinct from the school's internal demographic data.
4. Transparency in Timeliness
Federal data often operates on a 1-to-2 year lag. We always display the specific academic year for any given metric so users are aware of the data's freshness. Where possible, we augment slow-moving federal directories with faster-moving state accountability releases.